Automation and Standardisation of Seatbelt Routing for Crash Analysis: Additional studies on the influence of the initial shoulder belt position on ATDs and HBMs chest injury metrics in frontal collision simulations

Abstract

Crash simulations are a useful method to examine the effect of forces acting in road traffic crashes on vehicle occupants, in order to develop safety strategies based on this. The manual creation of seatbelts in a computational environment is not only very time-consuming, but also deviations in the seatbelt position arise due to the subjective judgment of an individual, lead to difficulties in comparing test results. This study aims to standardise and automate the process of seatbelt creation, significantly reducing the time input while ensuring reproducibility. The project was carried out using the Finite Element pre-processor ANSA from BETA CAE System. A Python-based automation for creating seatbelts has been developed, which creates a seatbelt within a few minutes for various combinations of vehicles and Anthropomorphic Test Device (ATD) models. The creation of these seatbelt models is reproducible. To evaluate how this unnatural perfection of the seatbelt position compares to a belt routed by a human before a crash test, the deviation in belt position was measured in repeated physical routings for different ATDs. The consequences of this natural lack of reproducibility was investigated by artificially recreating the deviating seatbelts with the automated routing tool and assessing the resulting biomechanical effects in frontal collision simulations. The developed automation tool has been observed to reduce the time required for creating a seatbelt model to just a few minutes. These generated belt models are reproducible and exhibit no visual variation in their initial belt position. The shoulder belt position can be adjusted based on the input parameter "Mouth to Belt distance" (M2B). The study revealed that the variation in M2B is influenced by the surface shape of ATDs. The convex shape of the Humanetics HIII 50th percentile male resulted in a 95mm difference between the lowest and highest M2B measurement. In contrast, the 5 th percentile female benefited from the natural belt guidance provided by the female chest, resulting in a difference of only 45mm between the lowest and highest M2B measurement. Furthermore, the examination of this parameter’s spread demonstrated an influence on the chest deflection of ATDs in frontal crash simulations, directly impacting the predicted occupant injury risk. In the case of Human Body Models, which are increasingly crucial in crash simulations, the study determined that the belt position notably affects rib strains and rib fracture probabilities across all age groups. The highest rib strain was observed in the region corresponding to the average position of the investigated belt routings.